Version: 2.0.0_devel (using unspecified) Compiler: mingw delivered with QtCreator 1.1.0 OS: Windows 32-Bit There seems to be a problem with curves created from equations. To reproduce: 1) plot Column 1 vs INDEX from gyrodata.dat 2) create equation [Column 1 (V2)]+2 and add in new plot 3) result: the new plot is empty, while the data vectors have been created correctly and the second curve as well, as seen from the data manager! Of course, it is expected that the curve be shown. Note that this is currently bothering us pretty badly. We'd be grateful if it could be fixed quickly, especially as it is a recent regression and probably relatively easy to fix.
I have just checked and it seems histograms and power spectra suffer from the exact same problem. Hopefully a common cause... I guess all plugins are impacted, which makes it pretty bad.
I am pretty puzzled. I have just done a clean build under Linux, and there it seems to work fine. I'll try again to rebuild a complete Windows version, but yesterday I already did it. Really strange! Is anyone else on seeing this on any platform?
I've been updating my machines here to svn trunk every morning and haven't seen this on Linux (checked only with PSDs which I use daily right now).
So, under Windows I still have the problem after a complete new build. I'll try it with a different Qt version from home, but it looks like it's platform- or version-dependent. I don't know if that's good news...
Unfortunately, on my home Windows machine with a different Qt version (4.6.1) and a debug build, I also have the problem. So it seems that there is a Windows-specific issue, but I really don't think it is linked with an unclean build. I suspect Matt is working under Linux exclusively. Barth, how about you: can you test/debug that under Windows? Or Peter, are you still around?
So, there we go: I have tracked down the problem to have appeared with revision 1054888 (the big update refactor). With revision 1051460 (the one just before) it works, and switching up just one revision (to 1054888), plus the patch from 1055898 needed to be able to use the data wizard with ASCII, it no longer works. Now the question is: why only under Windows???
On Windows PC, I also get the problem with the equation. However when playing long enough with the curves using the equation vectors, I managed to see the curves but it disappeared when stored in a .kst file. With the last revision (if I knew how to get the revision number), I also have some problems under linux. With an english Ubuntu everything works almost fine, but with a german Ubuntu (from a colleague) the equation's editor does not understand float constant: "* 2.1" will be understood (and stored) as "* 2" meaning also that "/ 0.1" will be understood as "/0" When we tried with "1,4" instead of "1.4" the value remains in the equation, but no value is computed...
The UpdateManager doesn't resolve dependencies, and the update order seems to be different on the platforms. In this case here the equation must first fill the vectors and update them which then will be used by the curve. Maybe there are other places with the same bug.
SVN commit 1106882 by netterfield: Found a serious bug in the update manager - previously, it only worked by chance.... I'm really hoping that this will fix the windows update bug. Can someone check? CCBUG: 225347 (also, fix a typo in data vector locking...) M +1 -1 datavector.cpp M +2 -2 updatemanager.cpp WebSVN link: http://websvn.kde.org/?view=rev&revision=1106882
See devel_docs/Kst2Specs/Updates for a description of the update strategy (which might work now...)
On Windows I could remove the work around, so your patch works.
I also confirm this seems to work OK now, so let's close it.
Change version from 2.0.0_devel to 2.0.0 to simplify version numbering.
These bugs are solved with 2.0.0